An Explanation of MPEG-2 for DVD Authoring

Kyle Cassidy
February 2006

» Next Image Previous Image «

"A full body cast?!" I shrieked.

"Well," my mother replied, "you know your great aunt Edna, she always said she was going to ski jump on her 80th birthday."

"She broke eight bones ski jumping?"

"Oh no, she went through the jump without a hitch. The broken bones came when she got thrown from the mechanical bull at the party afterwards."

And there she was, poor Aunt Edna, stretched out on a hospital bed wrapped in plaster from nose to toe. She wasn't able to do much of anything on her own, so my mother insisted I come visit with her.

"So," I said, sitting down, "is the food here any good?"

Aunt Edna's eyes darted left and right. She was always sharp as a tack. The clock ticked. Somewhere I heard a dog bark. I tried to think of something that was going on in my life that I could tell her about.

"Hey!" I said finally, "let me tell you about MPEG-2 encoding for DVDs!"

MPEG stands for Motion Picture Experts Group. As you might imagine, it's a consortium of people who know a lot about how to make moving pictures and wanted a digital compression standard which would best represent their interests. This means getting the highest quality of video in a relatively small space. There are two types of MPEG encoding that can be used on a standard video DVD: MPEG-1 and MPEG-2. MPEG-1 is almost never used because of its comparatively pathetic bitrate.

MPEG-2 is a highly compressed, lossy format. ("Lossy" means you "lose" a lot of image quality when you apply a lot of compression.) Different encoders produce different results, and some compress better than others do. How good your MPEG encoder is has a direct result on how much video of what quality you can fit on a DVD.

MPEG compresses groups of frames together. This is very useful -- if the sky is blue in frame one, chances are it will be blue in the next frame, and the one after that, and the one after that. The more repetition there is, the higher the rate of compression. If, for example, you had a 900 page document that consisted solely of the letter "a" replicated two million times, you could compress that entire book down to the instruction "repeat the letter 'a' two million times". We use a rudimentary form of compression every day when we use "Dr." to represent the word "Doctor" and "St." for "Street." Both of these we decompress on the fly without thinking of it when we say, "Doctor Jones lives on Elm Street." We have these automatic compression schemes because we use these words frequently. By viewing multiple frames, the MPEG compressor is able to find frequently used "words" and come up with the best compression for them.

By doing this, MPEG can achieve remarkable rates of compression. An hour of DV encoded at a rate of 5:1 can take up 20 gigabytes of disc space while a 4.37GB DVD can fit a two hour movie on it and still have space left over for a couple of trailers and a blooper reel. That's significant compression. The down side of this great compression is that one blob of data now represents several frames of video -- which means there is no frame-accurate editing without that blob of frames first being uncompressed. MPEG, remember, is lossy compression, so every time you uncompress and re-compress, you lose video quality.

• Digg This!
• del.icio.us
• Technorati
• StumbleUpon
• Reddit